Alpaca

You can run a dedicated MCP server to connect Alpaca’s trading capabilities with AI assistants. This server enables real-time market data access, order placement, and portfolio management through MCP-enabled clients like Claude Desktop, Cursor, and VS Code. It supports stocks, options, crypto, and comprehensive trading workflows, while letting you operate from local or remote setups.

How to use

Connect Alpaca MCP Server to your MCP client (for example Claude Desktop, Cursor, or VS Code). Start the local server for interactive use, or run a remote HTTP transport for access from another machine. Your MCP client will send natural language trading requests, and the server translates them into Alpaca API calls once you provide your API credentials. Make sure you configure your client with the proper API keys and trading mode (paper or live). You can place trades, query balances, inspect positions, fetch market data, and manage watchlists through natural language prompts.

How to install

Prerequisites: Python 3.10 or newer, and uv available on your system.

One-click local setup with uvx (recommended):

uvx alpaca-mcp-server init

If you don’t have uv yet, install it first and restart your terminal so uv/uvx are recognized.

Configure the MCP client with your Alpaca API keys in the client’s environment block (examples shown for JSON config).

Example minimal local config for the MCP server (stdio mode) to run with uvx and environment keys for Alpaca credentials:

{
  "mcpServers": {
    "alpaca": {
      "command": "uvx",
      "args": ["alpaca-mcp-server", "serve"],
      "env": {
        "ALPACA_API_KEY": "your_alpaca_api_key",
        "ALPACA_SECRET_KEY": "your_alpaca_secret_key"
      }
    }
  }
}

Additional setup for remote HTTP transport

To access the MCP server from another machine, run the server with HTTP transport and configure the client with the HTTP URL.

Start the server for remote access (default host and port):

alpaca-mcp-server serve --transport http

Or bind to a specific host and port for remote usage:

alpaca-mcp-server serve --transport http --host 0.0.0.0 --port 9000

Client configuration example (Claude Desktop / Cursor) using HTTP transport:

{
  "mcpServers": {
    "alpaca": {
      "type": "http",
      "url": "http://your-server-ip:8000/mcp",
      "env": {
        "ALPACA_API_KEY": "your_alpaca_api_key",
        "ALPACA_SECRET_KEY": "your_alpaca_secret_key"
      }
    }
  }
}

Available tools

get_account_info

View balance, buying power, and account status.

get_positions

List all held assets.

get_open_position

Detailed info on a specific position.

close_position

Close part or all of a position.

close_all_positions

Liquidate entire portfolio.

get_stock_quote

Real-time bid/ask quote for stocks.

get_stock_bars

Historical OHLCV bars with flexible timeframes.

get_stock_latest_trade

Latest market trade price.

get_stock_latest_bar

Most recent OHLC bar.

get_stock_snapshot

Comprehensive snapshot with quote, trade, and bars.

get_stock_trades

Trade-level history for stocks.

get_orders

Retrieve all or filtered orders.

place_stock_order

Place stock orders of various types.

cancel_order_by_id

Cancel a specific order.

cancel_all_orders

Cancel all open orders.

place_crypto_order

Place crypto orders with multiple time-in-force options.

get_option_contracts

Get option contracts with flexible filtering.

get_option_latest_quote

Latest quote for an option contract.

get_option_snapshot

Get Greeks and underlying for an option.

place_option_market_order

Execute option strategies.

exercise_options_position

Exercise a held option to convert to underlying.

get_market_clock

Market open/close schedule.

get_market_calendar

Holidays and trading days.

get_corporate_announcements

Historical and future corporate actions.

create_watchlist

Create a new watchlist.

update_watchlist

Modify an existing watchlist.

get_watchlists

Retrieve all saved watchlists.

get_asset_info

Search asset metadata.

get_all_assets

List all tradable instruments with filters.